Feedback technology isn’t as simple as it seems. Over the years, the technology has improved and costs have decreased dramatically. HP’s scribed disc encoder used to be the standard at around $50/unit in small quantities. Now through beam LED based encoder kits are available for $25 and less. Accuracy is good, and products are available from a variety of suppliers off the shelf.
There are still some caveats. There is a tendency to specify more precision that is actually needed just because its cheap and available. But there is a hidden pitfall to this approach. Too much feedback precision can actually turn into “jitter”. Since the controller can see high input rates of quadrature data and resolve them to 1 bit, the least mechanical error in the transmission system will result in encoder input fluctuations that don’t really represent actual motion in the load.
There aren’t any real rules about how to specify encoder precision for a given application. In control theory, you don’t want more than ten times the precision in the sensor compared to the desired accuracy at the load. So that’s a major boundary. But its not always considered, so this is a good place to exert some caution.
Generally, control systems are specified at 1 megahertz frequencies. And 4 megahertz is common on major platforms. So bandwidth is not a limiting factor. But when linear motors are used in semiconductor applications, the very high precision, typically 1 micron per pulse of feedback, quickly turns into higher bandwidth requirements. So some of the higher performance motion platforms run at frequencies of 10 megahertz and higher to handle these situations.
The interesting thing is that most motion applications don’t really require extreme precision. A typical requirement of 0.001″ of position is precise enough for the majority of applications. So it seems that the quest for ever more precision runs counter to the need in the market.
There are some interesting new devices coming Renishaw and others based on Hall effect magnetic feedback. This new class of devices is lower cost, easier to use and will produce feedback resolution up to 8192 pulses per revolution. The chip based solution presents some unique packaging constraints, but because the magnetic field is less sensitive to air gap and radial precision, its easy to adjust.
Magnetic feedback also has another advantage, its less sensitive to temperature. Some of the magnetic solutions are specified to operate up to 150 degrees Centigrade. given that we are generally using the feedback device on an electric motor, this is a good match of technology and application.
While high performance applications will continue to drive manufacturers to achiever greater precision, there is plenty of room for unique low resolution plaftorms. I think the future of feedback is going to very interesting indeed.
Tell Us What You Think!